Airbyte Adds Semantic Search

Airbyte, an open data movement platform provider, has added semantic search in its context layer, enabling artificial intelligence agents to find relevant information, such as customer conversations, documents, and support tickets, and combine those findings with structured data in Snowflake and BigQuery.

Available through the Airbyte Model Context Protocol (MCP) server, the capabilities let users ask business questions in natural language from Claude, ChatGPT and other MCP-compatible AI assistants.

In addition, Airbyte expanded its certified connector portfolio and introduced improvements to BigQuery, MySQL and Zoom data connectors. The BigQuery source connector reads 1 terabyte of data in less than an hour. The Airbyte context layer supports direct queries of Snowflake and BigQuery destinations containing data synced from more than 50 Airbyte source connectors. Semantic search adds the ability to search supported text sources by meaning, helping agents answer questions such as "Which customer accounts are the best prospects for upgrades?" using both business records and the context in customer interactions.

"An account record can tell you when a customer's contract renews. A call transcript can tell you what would convince them to expand," said Michel Tricot, co-founder and CEO of Airbyte, in a statement. "We're bringing those perspectives together so AI agents can find relevant evidence, connect it to business data, and help people act on what they learn."

Airbyte's semantic search retrieves information based on its meaning. Hybrid search capabilities combine semantic and keyword matching, preserving exact matches for terms such as product names and order numbers. This makes it possible to ask questions such as the following:

• Why did refunds increase last week, and which account plans were affected?
• Which accounts raised concerns on calls this quarter, and what are their current opportunity stages?
• Which support conversations describe slow performance following the latest update?
• Where did we define an active customer, and which definition is most recent?

Airbyte indexes selected text fields from data customers already replicate and then runs the embedding model on its own infrastructure, so text is not sent to a third-party embedding API.

Supported semantic search sources include Confluence, Gmail, Google Drive, Gong, Jira, Outreach, Pylon, Salesforce, and Zendesk Chat. Through the MCP server, agents can combine semantic search, warehouse SQL queries, and live reads from supported sources.

Airbyte also expanded its certified connector portfolio with Ashby, Gmail, MySQL, Pipedrive, and Reddit Ads. Airbyte engineers maintain these connectors, back them with support service-level agreements, and follow a published policy for breaking changes.

Airbyte also introduced its rebuilt BigQuery source connector, which reads one terabyte of data in less than an hour on Airbyte Cloud. The connector supports resuming interrupted syncs from the last checkpoint within BigQuery's read-session window, helping teams move large datasets into their chosen destinations.

Airbyte's Zoom source connector provides access to meeting and webinar records, participant reports, registration and poll data, webinar questions and answers, and Zoom Events. Teams can bring these records into their data platform to analyze attendance and engagement alongside customer and sales data. The Zoom connector supports OAuth authentication, incremental syncing for selected meeting, webinar, and reporting streams.